This file was created by scanning the printed publication. Errors identified by the software have been corrected; however, some errors may remain. Nurse Associations in the Chihuahuan Desert Shrublands

Jamey Thompson Laura F. Huenneke

Abstract-Spatial studies in the Chihuahuan Desert of associa­ influence on the soil temperature beneath it (Miller and tions with Larrea shrub islands found juveniles more often associ­ Stoner 1987). Evaporation and transpiration are decreased ated with shrub islands than unassociated. The spatial structure of at lower soil temperatures by the ameliorating effect of the the shrub islands points to Nurse Plant facilitation of seedlings. canopy (Tiedemann and Klemmedson 1973, 1977; Nobel Experiments tested the effects of canopy, shrub islands, and under­ 1980; Belsky 1994; Montana 1994). story on five perennials' germination. Longer survival times, but not At best, though, Nurse Plant relationships are com­ higher germination, of acerosa were found beneath artificial mensalistic for the seedling; at worst the seedlings compete canopies in intershrub areas and cleared shrub islands as compared with the adult plant for scarce resources. In arid zones, water to intact shrub islands. Therefore, Zinnia may be a poor competitor, stress can be the most limiting factor for , especially for preferring germination microsites with few competitors, regardless seedlings' shallow roots (Leishmann and Westoby 1994). The of Nurse Plant effects. shallow rooting structures of seedlings compete not only with other seedlings (Eldridge and others 1991; Aguiar and Sala 1994; Leishmann and Westoby 1994), but also with the shallow roots ofthe adult plant (Tiedemann and Klemmedson It is often suggested in arid and semiarid environments 1977; Aguiar and Sala 1992; Belsky 1994). that adult plants facilitate germination and establishment: Larrea dominate creosote (Larrea tridentata [D.C.] Cov.) This idea is commonly called the Nurse Plant Hypothesis communities in southern New , complemented by a (Went 1942; Nabhan 1975; Nobel 1980; McAuliffe 1984a; diverse, spatially associated mix of shrubs, succulents such ?arls~on and Callaghan 1991; Cody 1993). Many seedlings as prickly pear cacti (Opuntia spp.), and annuals. Perennial In and and other zones can be found associated either grasses often sparsely populate large intershrub areas, directly under (Nobel 1980) or nea"r the canopy of adult which have lower soil nutrient moisture levels as compared plants (Carlsson and Callaghan 1991; Aguiar and Sala 1992; to the shrub islands (Schlesinger and others 1990). The Belsky 1994). The mechanisms of these associations are climate is typical of the Chihuahuan Desert, with irregular often not known, but are usually credited to physical or biotic spring rains and seasonal storms in the summer (Conley and facilitation provided by the "safe sites" of the Nurse Plant as others 1992). compared to other microsites (Fowler 1986a, 1986b; The spatial structure of the Larrea shrub islands points to Huenneke and Sharitz 1986; Fowler 1988; Scherff and possible Nurse Plant facilitation of seedlings. To test this others 1994). These cited physical facilitations include' possibility, germination trials were conducted under a split­ increased nutrient levels (Vetaas 1992; Belsky 1994) de~ plot design to test the effects of 1) canopies 2) shrub islands cre?sed soil temperatures (Nobel 1980), and increased soil and 3) understory on germination. Main factors that influ­ mOIsture (Nobel 1980; Belsky 1994), all compared to areas ?nce germination include light, water, and nutrients. Light outside of the Nurse Plant's influence. IS generally not considered a limiting factor in the desert On shrub islands in arid zones, the gradients of soil especially under the open canopies of Larrea. If the mai~ parameters such as carbon, nitrogen, and moisture shift limitations to germination and establishment are soil tem­ dramatically across centimeters (Charley 1972; Charley and peratures and nutrient levels, then an artificial shrub canopy West 1975; Jackson and Caldwell 1993). The microsites or increased nutrient loads respectively should stimulate provided by these putative Nurse Plants have higher nutri­ seeds to germinate. However, if the main limitation to ent levels (Vetaas 1992), translating into higher productiv­ germination and establishment is soil moisture, then the ityfor associated plants (Tiedemann and Klemmedson 1973' absence of other seedlings should increase germination and 1977; Kellman 1979; Escudero and others 1985; Georgiadi~ establishment. 1989). These nutrient differences are primarily limited to the top ofthe soil horizon where adult roots are rare (Charley and West 1975; Freckman and Virginia 1989; Franco and Materials and Methods ------others 1994). The Nurse Plant's canopy also has a strong Site Description

I?: Barrow, Jerry R.; McArthur, E. Durant; Sosebee, Ronald E.; Tausch, The experiments were conducted on the State Robm.J., co~ps. 1996. Proceedings: shrubland ecosystem dynamics in a University College Ranch near the Jornada Experimental changIng envlI'Onment; 1995 May 23-25; Las Cruces, NM. Gen. Tech. Rep. INT-GTR-338. Ogden, UT: U.S. Department of Agriculture Forest Service Range, 37 km NNE of Las Cruces, New Mexico. Creosote Intermountain Research Station. " (Larrea tridentata [D.C.] Cov.) and mesquite (Prosopis ~amey Thomp.son is Conservation Analyst, Green Mountains Restoration glandulosa Torr.) dominate the shrub community, with a Pro~ect, 236 Mam Street, Burlington, VT 05401. Laura F. Huenneke is Asslstant Professor, Department of Biology, New Mexico State University mix of perennial shrubs, subshrubs, cacti, and herbaceous Las Cruces, NM 88003-0001. '

158 annuals making up the rest of the species. The site is on an coated with Tree Tanglefoot™, a commercial resin that alluvial wash of the Dona Ana Mountains, with alkaline soils traps insects. To exclude the kangaroo rats and bird preda­ and a shallow calcic horizon (Buffington and Herbel 1965). tors, 50 em high, I-em diameter chicken-wire exclosures Mean summer temperatures approach 40 C (Conley and surrounded the seed cages. Finally, to minimize seed move­ others 1992), but during the study the June/July mean ment due to water or wind, a 1 cm thick layer of common sand maximum temperature was near 43 C. Mean precipitation is was spread over the plot after the seeds were dispersed. 218 mm, with the bulk (> 120 mm) coming in strong thunder­ The plots were monitored every three days from July 6th, storms in July, August and September, that begin July 4th 1994, until September 12th, 1994, and thereafter sampled ±14 days (Buffington and Herbel 1965). During the study every 21 days until early January 1995. A plastic overlay for summer the summer rains began on July 26th and brought each replicate allowed the location and survival of each 32 mm of rain at the germination site during July and seedling to be recorded. Data from these overlays were then August. translated into the survival curves for each species. Microsite resource levels were measured for soil moisture and soil temperature. Soil moisture was measured gravi­ Shrub Island and Intershrub Area Criteria metrically once a week for 8 weeks from the time the seeds The Larrea shrub islands that were chosen were mature were sown on July 6, 1994 until late August 1994. Ten gram enough to have few shallow roots (> 1.5 meters tall) (Franco subsamples from the plots were dried until the weight varied and others 1994), and at least 2 meters in diameter. less than 0.003 g from day to day. Soil temperature was Intershrub areas were paired with chosen shrub islands and measured using a thermocouple (Omega II) on six occasions were at least 4 meters in diameter and had all grass clumps (6 a.m. on July 14th, 1994, 12 p.m. on July 12th, 1994,4 p.m. andjuvenile perennial plants removed. From the sample, we on July 6th, 1994, and 8 a.m., 12 p.m., and 4 p.m. on chose a random sub-sample of 48 intershrub areas and 48 September 20th, 1994). Since the soil was variable across shrub islands. The sample of randomly chosen shrub each plot, temperature readings were taken at a depth of islands was nearly uniform in size and understory species 10 cm in the northwest and southeast comers in each plot composition. and then averaged for the plot. The temperature, moisture, and germination data were then compared with one-way GLM models using repeated Shelters and Seed Cages measures analysis and contrasts (SAS 1994). Only Zinnia germinated, but had a non-normal distribution ofgerminants Two experiments on seedling germination were conducted. (but not survival of germinants). Tukey's transformation The first experiment was a factorial of four treatments with (sqrt[n + sqrt(n)]) shifted the data to normality. The analy­ eight replicates each testing the effects of shrub islands and ses used one-way GLM models broken into three compari­ intershrub areas on germination and establishment. In this sons. First, the four treatment factorial was divided into factorial, all shrub islands were treated by removing all presence/absence of shrub island and presence/absence ofan aboveground shrub and plant biomass on the shrub island. artificial shade canopy and were then compared using re­ All intershrub areas were treated by removing all grasses peated measures analyses. A second series of comparisons and other plants within one meter of the center of the plot. used one-way comparisons of artificial (lSA-Shade, SHR­ Artificial canopies sheltered half of the shrub islands (SHR­ Shade) and natural canopies (No US, SHR-Intact) and Shade) and intershrub areas (lSA-Shade) while the other thirdly, understory (No US) and intact shrub islands (SHR­ two treatments of the factorial were left uncovered (SHR­ Intact). Open, ISA-Open). To simulate the shrub canopy, these shade canopies were constructed from 1 m2 60% shade cloth (approximately the shade percentage of adult Larrea; V. Results Gutschick, pers. comm.) and held in place one meter above ------the plots. A second experiment of two treatments tested the Resource Levels effect of understory on germination. In the first treatment, the shrub canopy was left intact and all understory biomass Mean moisture content varied among treatments with the was removed (No US); the other treatment was left entirely artificial shade treatments (ISA-Shade and SHR-Shade) intact and served as the control for shrub islands in both having higher values than the naturally shaded treatments experiments (SHR-Intact). (table 1). These artificial shade treatments provided more The experimental unit in both experiments was the seed soil moisture for potential germinants than these other cage. To test the differences in the microsites, seeds of five treatments across the whole study. There was no difference species (Zinnia acerosa (100 seeds), Larrea tridentata (50 in terms of soil moisture between intershrub areas and seeds), Prosopis glandulosa (30 seeds), Gutierrezia micro­ shrub islands or between artificial canopy and open treat­ cephala (50 seeds), and Opuntia imbricata (50 seeds» were ments. Understory plants did increase soil moisture ofshrub dispersed on July 6th, 1994, in each 100 cm2 seed cages. islands, but not significantly. Larrea, Prosopis, and Opuntia were all scarified, either by Time had a significant effect on soil moisture, most likely acid or mechanically, before they were placed in plots. due to the lack of rainfall at the site. The only precipitation Three major threats exist for seeds in the desert: ants, came during week four when the site received 32 mm of rain. kangaroo rats, and wind or water movement. To exclude ants, The moisture means for week four show that there were an exclosure of 15 cm garden fencing formed a rectangle of higher levels of soil moisture with artificial shade canopies 2000 cm2 that was centered on the seed cage. The edging was on week four compared to natural canopies (table 1, fig. 1).

159 Table 1-The effects of soil moisture on treatment. See text for details on comparison methods.

Time Mean examined Comparison df square Fvalue P

0.08,------_---, Artificial shade treatments had higher soil moistures than .~ 0.07+------1------1 unshaded treatments, but shrub islands were not different ~ 0.06 +---.;-----.---­ o than intershrub areas, eliminating any retention effect by N :r: 0.05 +---1---'--- the higher levels of organic matter on the shrub islands. The ~ 0.04 interaction term of shade*shrub was significant, indicating ~ .E 0.03 that the combination of shade treatment with the absence of III a shrub island (lSA-Shade) resulted in higher soil moisture '0 0.02 :E than SHR-Open (fig. 1). There were positive effects on soil '0 om CI) moisture of removing understory biomass, but overall, the o artificial canopies produced the strongest effect on moisture SHR-Shade SHR-Open ISA-Shade ISA-Open No US SHR-Intact levels. Treatment Mean soil temperature did not vary among treatments as natural canopies and artificial canopies were not different, Figure 1-Soil moisture means after 32 mm of rain nor were there any effects of the understory on soil tempera­ during the fourth week of the study. ture (table 2). There were significant differences in soil temperature across time, primarily due to the inclusion of the July afternoon data. There was also an interaction of

Table 2-The effects of treatment on soil temperature. Comparisons follow method described in text.

Time Mean examined Comparison df square Fvalue P

Entire study Time 5 42306.920 715.98 0.0001**** Time*Shade 5 1076.718 18.22 0.0001**** Time*Shrub 5 26.679 0.45 0.8117 Time*Shade*Shrub 5 45.502 0.77 0.5726 Contrast: Natural canopy! 1 78.973 0.35 0.5529 Artificial canopy Contrast: No understory! 2.870 0.01 0.9127 Intact shrub is!. July afternoon Shade 1278.443 46.59 0.0001**** Shrub 37.719 1.37 0.2513 Shade*Shrub 24.462 0.88 0.3554 Contrast: Natural canopy! 1178.032 40.71 0.0001 **** Artificial canopy Contrast: No understory! 11.390 1.19 0.2943 Intact shrub is!.

160 the artificial canopy treatments and time (shade*time) caused presence of an artificial canopy cooled the soil more than a by the artificial canopy treatments were cooler over the natural canopy did. whole study. Artificial canopy and open treatment did not differ, nor were there any differences between shrub islands and intershrub areas in terms of soil temperature across the Germination Trials whole study (table 2). For Zinnia, the survival data were normal, but the num­ The average mean temperatures of each treatment through ber of germinants was not normal, so the number of ger­ the different time periods show no difference except for the minants was transformed using Tukey's transformation 4 p.m. readings in early July (fig. 2). The mean temperatures (fig. 3). The results on the transformed germinant number of the treatments on July 6th at 4 p.m. show higher values show no differences (table 3). The Shrub/Shade comparison for the artificial shade treatments (lSA-Shade and SHR­ found an interaction of shrub*shade indicating that shrubs Shade) as compared to both unshaded and natural canopy without canopies and intershrub areas with canopies had treatments (table 2). These two artificial canopy treatments high germinant numbers. Survival was similar among treat­ were cooler than the rest of the treatments. The presence of ments except for ISA-Shade (table 3, fig. 4). Germinants an understory had no effect on soil temperature for shrub survived for a longer period of time in intershrub areas than islands. Shrub islands were not different from intershrub on shrub islands, due primarily to the effects of the artificial areas, demonstrating that the results for soil temperature canopy. This indicates that the lack of competition from are the same as the results of the moisture data. The juvenile and adult plants was the most important factor for survival and establishment.

0 60

Table 3-The effects of treatment on transformed germinant number and on survival. See text for details on comparison method.

Mean Type Comparison square Fvalue P

161 45 • • ... SHR-Open T SHR-Shade 40 o SHR-Intact • • No Understory 35 o ISA-Open ...... • • ISA-Shade ... 30 • ~ 0 > ... 25 ·e:s 'Y 'Y ... CI) - ..... T 0... • • ... ~ 20 • .0 a a i T ... :se T Z (") 15 u • T • 0 ... T T 10 rio • 0 • ... - T a 0 • • T ... 0 0 rio • • - 0 c- a • • • ~ i o - - 3 6 12 15 18 21- 24- 27- 30- 33 36 Time in days

Figure 4-Survivorship curve for Zinnia germinants.

Discussion evaporating the upper layers of the soil horizon. Previous ------work by Schlesinger and others (1990) measured soil mois­ Microsites and Resources ture differences between Larrea shrub islands in intershrub areas and found a similar canopy shading effect on soil Subtle environmental differences in light and water moisture. Belsky (1994) suggests that Nurse Plant effects strongly influence the gennination of seeds in arid environ­ should be more noticeable in more arid environments as the ments. Harper and others (1965) suggest that germination effects of shade on water relations become stronger. and early establishment may have the strictest require­ The temperature data suggest that the most significant ments of all the lifestages of the plant in tenns of nutrients factor for soil temperature in different microsites could be and competition. The most important factors affecting soil the quality and density of the canopy shading the soil. For all moisture levels in deserts are root density (Fenner 1985; time periods, both early and late in the summer growing Franco and Nobel 1988) and soil temperature (Nobel 1984). season, the only difference in soil temperatures among Microsites that are exposed to direct solar input and those treatments came at the daily high temperature early in the with shallowly rooted plants are less favorable to an arid season (4 p.m. on July 6th). At this time, the artificial canopy environment seedling, since increased evaporation lowers treatments (lSA-Shade and SHR-Shade) were 10-15 Clower soil moisture. Mesquite (Prosopis) and creosote (Larrea) are compared to the other treatments. This points to the possi­ both dominant shrubs with strong shrub island associations bility that the artificial canopies provided a different type of in many Chihuahuan Desert communities, but both have shading than did natural canopies. The open Larrea canopy relatively open canopies, limiting possible Nurse Plant fa­ pennits larger amounts of direct sunlight compared to the cilitation (Bush and van Auken 1990). more diffuse shadecloth, even at the same 60% total shading. This difference could eliminate comparison with a natural Larrea canopy, since the soil would receive alternate periods Do Canopies Affect Soil Moisture and of direct sun and sharp shadow. Temperature? If the artificial canopy treatments provided more canopy shade than did natural ones, this points to a preferred The artificial canopy treatments had higher levels of soil microsite having the most canopy between the soil and the moisture compared to shrub canopies, since both ISA-Shade sun in times of peak temperatures. Larrea shrub canopies and SHR-Shade had significantly higher soil moisture levels have never been studied for effects on gennination, but some over the whole study compared to the unshaded (lSA-Open argue that Larrea's canopy is too open to provide any and SHR-Open) and the naturally shaded treatments (No substantive Nurse Plant effect on soil temperature. Juvenile US and SHR-Intact). More moisture is available without columnar cacti in the associate with the competition for root uptake and without direct solar input nurse plant Hilaria rigida, a perennial grass which provides

162 a small, dense canopy (Nobel 1980; Franco and Nobel 1988; suggests that Zinnia's survival may depend on a lack of Cody 1993). The current data support McAuliffe, since the competitors. Thus, the optimal association for establish­ temperatures ofthe intact shrub (SHR-Intact) are not differ­ ment of a Zinnia is then a Larrea shrub island with few ent than those of the intershrub areas (ISA-Open). understory plants. The germination study points to a variety of mechanisms that could account for the moisture and temperature data. Do Shrub Islands Influence Germination? The most obvious is a physical facilitation of soil factors. It There were no significant differences in germinant num­ is possible that there was some facilitation of soil factors by ber among the treatments, but mean survival was lower for shrub islands such as more available moisture that gravi­ shrub islands compared to intershrub areas. In ISA-Shade, metric methods would not detect. It is also possible that any Zinnia survived longer with the higher moisture levels and historic Larrea shrub islands' facilitation of Zinnia went lower temperatures than in the conditions of any other undetected during the study due to unusual weather. Fi­ treatment. With the data for soil moisture and temperature nally, it is possible that the observed Zinnia-Larrea spatial included, this points to two possibilities: lower moisture and associations represent germination under different soil con­ higher temperatures favor Zinnia germination or Zinnia is ditions (for example, higher moisture, lower soil tempera­ a poor competitor and favors more depauperate areas. The tures) and consequent juvenile survival that is not depen­ data refute the first possibility since none of the summer's dent on shrub islands. A more complete picture of species' germinants survived longer than 36 days in any treatment. responses would involve examining the spatial association As for the second possibility, Zinnia is a poor competitor for of Zinnia and understory plants around Larrea shrub is­ nutrients and water since ISA-Shade dominate the survival lands during a summer with a more regular rainy season. data compared to the shrub island treatments. Last sum­ mer, Zinnia survived longer in the more resource-limited, References depauperate intershrub areas. It is possible that Zinnia's ------preferred microsite is a young Larrea shrub island without Aguiar, M. R. and O. E. Sala. 1992. Competition and facilitation in a substantial understory. Thus, aLarrea shrub island would the recruitment of grass seedlings in Patagonia. Funct. Ecol. 6: provide adequate shade for the germinants, lower soil tem­ 66-70. Aguiar, M. R. and O. E. Sala. 1994. Competition, facilitation, seed peratures through shading, but higher soil moistures in the distribution and the origin of patches in a Patagonian steppe. absence of understory competitors. Oikos 70: 26-34. Belsky, A. J. 1994. Influences of trees on savanna productivity: Tests of shade, nutrients, and tree-grass competition. Ecol. 75: Are There Nurse Plants in the Chihuahuan 922-932. Buffington, L. C. and C. H. Herbel. 1965. Long term rain patterns Desert? on the Jornada Experimental Range. Ecol. Monogr. 35: 139-164. Bush, J. K. and O. W. van Auken. 1990. Growth and survival of There have been several studies of nurse plants in deserts Prosopis glandulosa seedlings associated with shade and herba­ ofthe Southwestern US and Mexico (Nabhan 1975; McAuliffe ceous competition. Bot. Gaz. 151: 234-239. 1984b; Yeaton and Manzares 1986; Franco and Nobel 1988; Carlsson, B. A. and T. V. Callaghan. 1991. Positive plant associa­ Valiente-Banuet and Ezcurra 1991; Cody 1993), but all of tions in tundra vegetation and the importance of shelter. J. Ecol. 79: 973-983. them focused on succulents in the Sonoran and Mojave Charley, J. L. 1972. The role of shrubs in nutrient cycling. in Desert. Shrubs and subshrubs in southern Mexico form Wildland Shrubs: Their Biology and Utilization. (Ed. by C. M. shrub islands that structure the community spatially McKell, J. P. Blaisdell, and J. R. Goodin), pp. 182-203. USDA (Silvertown and Wilson 1994). Spatial studies in the South­ Forest Service Gen. Tech. Report INT-1. west US and Mexico have found that there is a significant Charley, J. L. and N. E. West. 1975. Plant-induced soil chemical patterns in some shrub-dominated semi-desert ecosystems of focusing of productivity around shrub islands along with a . J. Ecol. 63: 945-963. range of interactions, from facilitation (nurse plants) to Cody, M. 1993. Do Cholla Cacti (Opuntia spp., Subgenus competitive inhibition (Charley and West 1975; Silvertown Cylindropuntia) use or need nurse plants in the Mojave Desert? and Wilson 1994). Silvertown and Wilson (1994) concluded J. Arid Env. 24: 139-154. Conley, W., M. R. Conley, and T. R. Karl. 1992. A computational that Larrea served as a "focus shrub" with all other species study of episodic events and historical context in long-term occurring in association with Larrea in the southern ecological processes: Climate and grazing in the northern Chihuahuan Desert. Other studies in the northern Chihuahuan Desert. Coenoses 7: 55-60. Chihuahuan Desert suggest that there are similar patterns Eldridge, D. J., M. Westoby, and K. G. Holbrook. 1991. Soil-surface for annuals focusing on Larrea and Prosopis in shrub com­ characteristics, microtopography, and proximity to mature shrubs: Effects on survival of several cohorts of Atriplex vesicaria seed­ munities (Lightfoot 1991). lings. J. Ecol. 78: 357-364. Our results show a similar pattern of association. These Escudero, A. B., B. Garcia, J. M. Gomez, and E. Luis. 1985. The results demonstrate only the shrub islands' effects impact­ nutrient cycling in Quercus rotundifolia and Quercus pyrenaica ing one species, Zinnia, and not surprisingly, suggest that it ecosystems ("dehasas") of Spain. Acta Oecologia, Oecological Plantarum 6: 73-86. is a complex mechanism that determines the regeneration Fenner, M. 1985. Seed Ecology. Chapman and Hall, New York, NY. niche of that seedling (Grubb 1979). According to spatial Fowler, N. S. 1986a. The role of competition in plant communities studies, Zinnia'.associated with a Larrea shrub island more in arid and semi-arid regions. Ann. Rev. Ecol. Syst. 17: 89-110. often than chance would allow. However, according to germi­ Fowler, N. S. 1986b. Microsite requirements for germination and nation and resource data here, Zinnia seedlings survived establishment of three grass species. Am. Mid. Nat. 115: 131-145. Fowler, N. S.1988. What is a safe site: Neighbor, litter, germination longer in intershrub treatments with higher moisture, date and patch effects. Ecology 69: 947-961. lower temperatures, and presumably fewer nutrients. This

163 Franco, A. C. and P. S. Nobel. 1988. Interactions between seedlings Montana, C. 1994. The colonization of bare areas in two-phase of Agave deserti and the nurse plant Hilaria rigida. Ecol. 69: mosaics of an arid ecosystem. J. Ecol. 80: 315-327. 1731-1740. Nabhan, G. 1975. Nurse Plant ecology of threatened desert plants. Franco, A. C., A. G. de Soyza, R. A. Virginia, J. F. Reynolds, and W. in Conservation and management ofrare and endangered plants: G. Whitford. 1994. Effects of plant size and water relations on gas Proceedings of a California Conference on the Conservation and exchange and growth of the desert shrub Larrea tridentata. Management of Rare and Endangered Plants. (Edited by T. S. Oecol. 97: 171-178. Elias) pp. 377-383. California Native Plant Society, Sacramento, Freckman, D. W. and R. A. Virginia. 1989. Plant-feeding nematodes CA. in deep-rooting desert ecosystems. Ecol. 70: 1665-1678. Nobel, P. S. 1980. Morphology, nurse plants, and minimum apical Georgiadis, N. J. 1989. Microhabitat variation in an Mrican sa­ temperatures for young Carnegeia gigantea. Bot. Gaz. 141: vanna: Effects of woody cover and herbivores in Kenya. J. Trop. 188-191. Ecol. 5: 93-108. Nobel, P. S.1984. Extreme temperatures and thermal tolerances for Grubb, J. P. 1977. The maintenance of species richness in plant seedlings of desert succulents. Oecol. 62: 310-317. communities: the importance of the regeneration niche. BioI. Rev. SAS. 1994. SAS User's Guide. SAS Institute, Cary, North Carolina. 52: 107-145. Scherff, E. J., C. Galen, and M. L. Stanton. 1994. Seed dispersal, Harper, J. L., J. T. Williams, and G. R. Sagar. 1965. The behavior of seedling survival and habitat affinity in a snowbed plant: limits seeds in soil: I. The heterogeneity of soil surfaces and its role in to the distribution of the snow buttercup, Ranunculus adoneus. determining the establishment of plants from seed. J. Ecol. 53: Oikos 69: 405-413. 273-286. Schlesinger, W. L., J. F. Reynolds, G. L. Cunnigham, L. F. Huenneke, Huenneke, L. F. and R. C. Sharitz. 1986. Microsite abundance and W. M. Jarrell, R. A. Virginia, and W. G. Whitford. 1990. Biological distribution of woody seedlings in a South Carolina cypress­ feedbacks in global desertification. Science 247: 1043-1048. tupelo swamp. Am. MidI. Nat. 115: 328-335. Silvertown, J. and J. B. Wilson. 1994. Community structure in a Jackson, R. B. and M. M. Caldwell. 1993. Geostatistical patterns of desert perennial community. Ecol. 75: 409-417. soil heterogeneity around individual perennial plants. J. Ecol. Tiedemann, A. R. and J. O. Klemmedson. 1973. Effects of mesquite 81: 683-692. on physical and chemical properties of the soil. J. Range Man. 26: Kellman, M. 1979. Soil enrichment by neotropical savanna trees. 27-29. J. Ecol. 67: 565-577. Tiedemann, A. R. and J. O. Klemmedson. 1977. Effect of mesquite Leishmann, M. R. and M. Westoby. 1994. The role of seed size in trees on vegetation and soils in the desert grasslands. J. Range seedling establishment in dry soil conditions-experimental evi­ Man. 30: 361-367. dence from semi-arid species. J. Ecol. 82: 249-258. Valiente-Banuet, A. and E. Ezcurra. 1991. Shade as the cause of Lightfoot, K. 199t. Associations of annual plants and shrubs in the association between the cactus Neobuxbaumia tetetzo and the northern Chihuahuan Desert. Master's Thesis. New Mexico State nurse plant Mimosa luisana in the Tehuacan Valley, Mexico. University, Las Cruces, NM. J. Ecol. 79: 961-971. McAuliffe, J. M. 1984a. Prey refugia and the distributions of two Vetaas, O. R. 1992. Micro-site effects of trees and shrubs in dry Sonoran Desert cacti. Oecol. 64: 82-85. savannas. J. Veg. Sci. 3: 337-344. McAuliffe, J. M. 1984b. Sahauro-nurse tree associations in the Went, F.W. 1942. The dependence of certain annual plants on Sonoran Desert: competitive effects' of sahuaros. Oecol. 64: annual shrubs in southern California deserts. Bull. Torr. Bot. 319-321. Club 69: 100-114. Miller, P. C. and W. A. Stoner. 1987. Canopy structure and environ­ Yeaton, R. I. and A. R. Manzares. Organization of vegetation mental interactions. in Topics in plant population biology (Ed. T. mosaics in the Acacia Schaffneri-Opuntia streptacantha associa­ S. Solbrig and others) pp.428-458. Leishman: Leipzig, Denmark. tion, southern Chihuahuan Desert, Mexico. J. Ecol. 74: 211-217.

164